Differential Sensitivities of Fast- and Slow-Cycling Cancer Cells to Inosine Monophosphate Dehydrogenase 2 Inhibition by Mycophenolic Acid
- 17 Downloads
As uncontrolled cell proliferation requires nucleotide biosynthesis, inhibiting enzymes that mediate nucleotide biosynthesis constitutes a rational approach to the management of oncological diseases. In practice, however, results of this strategy are mixed and thus elucidation of the mechanisms by which cancer cells evade the effect of nucleotide biosynthesis restriction is urgently needed. Here we explored the notion that intrinsic differences in cancer cell cycle velocity are important in the resistance toward inhibition of inosine monophosphate dehydrogenase (IMPDH) by mycophenolic acid (MPA). In short-term experiments, MPA treatment of fast-growing cancer cells effectively elicited G0/G1 arrest and provoked apoptosis, thus inhibiting cell proliferation and colony formation. Forced expression of a mutated IMPDH2, lacking a binding site for MPA but retaining enzymatic activity, resulted in complete resistance of cancer cells to MPA. In nude mice subcutaneously engrafted with HeLa cells, MPA moderately delayed tumor formation by inhibiting cell proliferation and inducing apoptosis. Importantly, we developed a lentiviral vector-based Tet-on label-retaining system that enables to identify, isolate and functionally characterize slow-cycling or so-called label-retaining cells (LRCs) in vitro and in vivo. We surprisingly found the presence of LRCs in fast-growing tumors. LRCs were superior in colony formation, tumor initiation and resistance to MPA as compared with fast-cycling cells. Thus, the slow-cycling compartment of cancer seems predominantly responsible for resistance to MPA.
Uncontrolled cell proliferation resulting from cell cycling deregulation is a hallmark of cancer. Although aggressive cancers are diverse and heterogeneous, they almost universally contain a fast-cycling compartment that can rapidly complete a cell cycle, and these cells are primarily responsible for the increase in tumor mass (1). This impressive proliferative capacity is, however, dependent on adequate supply of nucleotides. Cellular nucleotide synthesis is biochemically complex, but requires various enzymes that can be targeted clinically, including inosine monophosphate dehydrogenase (IMPDH), which is a rate-limiting enzyme in de novo synthesis of guanine. The enzymatic activity of IMPDH is composed of two separate isoforms, type 1 and 2 (2). The IMPDH2 isoform is associated with aggressive cancerous disease in experimental cancer (3, 4, 5, 6), and related to poor survival in osteosarcoma patients (7). Mycophenolic acid (MPA) acts as a nonnucleoside, noncompetitive, reversible inhibitor of IMPDH with fivefold higher potency of inhibiting IMPDH2 than IMPDH1. It has been reported to be able to inhibit cancer cell proliferation and induce apoptosis in several experimental models of human solid tumors and hematological malignancies by depleting guanine nucleotide pools (5,8, 9, 10).
In the last decade, interest into the importance of cell cycle velocity heterogeneity of cancers has increased. Although it was initially thought that cancer cells universally cycle and grow faster than normal cells, recently a slow-cycling (largely quiescent) compartment, which does not divide frequently but has the capacity to generate progeny that can repopulate the fast cycling compartment, has been identified in many tumors, (11). Functionally, these slow-cycling cancer cells appear to be associated with the capacity to generate new metastases while having superior resistance to therapy (12). Technically, these slow-cycling cells are identified by their capacity to retain a pulse label as faster cycling cells lose the pulse label at cell division. Thus these cells are indicated with the term label-retaining cells (LRCs) (13).
In this study, we aim to develop a lentiviral vector-based Tet-on label-retaining system that allows us to identify slow-cycling cancer cells in vivo, so that they can be subsequently isolated for functional characterization. We exploit this system to investigate the different sensitivity between fast- and slow-cycling cancer cells to IMPDH2 inhibition by MPA.
Materials and Methods
Stocks of MPA (AMRESCO LLC) were dissolved in dimethyl sulfoxide (DMSO) (Sigma-Aldrich). Doxycycline, collagenase IV and DNase were purchased from Sigma-Aldrich Corporation. Antibodies against IMPDH2, p-Histone3 and cleaved caspase-3 were purchased from Abcam Company, Millipore Corporation and Cell Signaling Technology, respectively.
To investigate the effects of MPA on cancer cells, seven different cancer cell lines derived from various tumor types were cultured. Human hepatoma cell lines HepG2 and HuH7, colon adenocarcinoma cell line Caco2, the epithelioid cervix carcinoma cell line HeLa and ovary adenocarcinoma cell line SKOV-3 were grown in Dulbecco’s modified Eagle medium (DMEM) (GIBCO, Life Technologies). Pancreatic cancer cell lines BxPC3 and PANC-1 were cultured with RMIP-1640. Both of the mediums were supplemented with 10% (v/v) fetal bovine serum (FBS) (Hyclone Technologies), 100 units/mL of penicillin and 100 µg/mL of streptomycin. All the cells were incubated at 37°C in a humidified atmosphere containing 5% CO2. All the cell lines were confirmed mycoplasma free and their STR genotyping was analyzed at the Department of Pathology, Erasmus Medical Center Rotterdam (Supplementary Table S1).
Lentiviral Vector-Based Tet-On Label-Retaining System
Lentiviral backbone plasmids pLV. EX3D/EF1A-rtTA (M2)-dsRed-Express2 and pLV.EX2D/TRE-eGFP were used to pack third generation lentiviral vectors (Supplementary Figure S1A). HeLa cells were transduced with both vectors to generate a system (HeLa Tet-on) that can express a histone 2B-green fluorescent fusion protein (histone-GFP) upon induction by doxycycline. GFP expression in vitro was analyzed from wk 0 to wk 3 by flow cytometry analysis (FACS) and confocal microscope (Zeiss LSM 510) (Supplementary Figures S1B,C). ZenlightEdition software was used to analyze confocal microscope images. Cells that maintained GFP expression over this period of time were identified as LRCs.
For identification of LRCs in vivo, female NOG mice at the age of 8 to 10 wks were purchased from Taconic Biosciences (Denmark). Animal experiments were performed with the approval of the institutional animal ethics committee (Dier Experimenten Commissie, Erasmus MC). Mice were bred in a special pathogen free (SPF) environment during the whole experimental period. Mice were injected subcutaneously with 5 × 106 HeLa Tet-on cells. After engraftment (10 to 15 d), water containing 1 mg/mL doxycycline and 5% sucrose was given for 5 d. Mice were killed at different time points after withdrawal of doxycycline. A portion of harvested tumor samples was fixed with 4% paraformaldehyde and embedded in paraffin for subsequent immunohistochemistry, while the remaining tumor tissue was dissociated with 5 mg/mL collagenase IV and 2 mg/mL DNase at 37°C for 30 min to obtain single cell suspension. Cells were further sorted as singlets for separation into non-LRCs (GFPlowdsRed) and LRCs (GFPhighdsRed) by FACS sorter (Supplementary Figure S2). non-LRCs and LRCs were injected subcutaneously (either 1,000 cells or 10,000 cells per injection as appropriate) on four sites in the mice. At the same time, two populations of sorted cells were plated for colony-forming unit (CFU) assay (treated with or without MPA).
Colony Formation Assay
Cells were harvested and suspended in culture medium, yields were quantified through counting and plated in 6-well plates (500 cells/well), and then treated with serial dilutions of MPA (1, 2, 3, 4 and 5 µg/mL). The control group was supplemented with an equal volume of PBS. For the cells derived from xenograft tumor, cells were seeded into 12-well collagen coating plates and cultured in medium with or without MPA (10 µg/mL). Formed colonies were fixed by 70% ethanol and counterstained with hematoxylin and eosin after two weeks. Colony numbers were counted and their sizes were measured microscopically through digital image analysis.
Cells were seeded in a 96-well plate at a concentration of 5 × 103 cells/well in 100 µL medium. Cells were incubated overnight to attach to the bottom of the wells, and then treated with serials dilutions of MPA (1, 5, 10, 15, 20, 25 and 30 µg/mL). Cell viability was analyzed by adding 5 mg/mL MTT (Sigma-Aldrich) and 150 µL DMSO. Absorbance was determined using Enzyme mark instrument at the wavelength of 490 nm.
Analysis of Cell Cycle
Cells (5 × 105/well) were plated in six-well plates and allowed to attach overnight, followed by an application of MPA at concentrations of 5, 10, 15, 20 and 25 µg/mL for 48 h. Vehicle control was performed through the addition of an equal volume of PBS. After 48 h, control and treated cells were trypsinized and washed with PBS and then fixed in cold 70% ethanol overnight at 4°C. The cells were washed twice with PBS and incubated with 20 µg/mL RNase at 37°C for 30 min, and then with 50 µg/mL propidium iodide (PI) at 4°C for 30 min. The samples were analyzed immediately by FACS. Cell cycle was analyzed by ModFit LT 3.0 software.
Analysis of Cell Apoptosis
Cell apoptosis analysis was performed by staining cells with annexin V-FITC and PI. Cells (5 × 105/well) were seeded into six-well plates and incubated at 37°C in 5% CO2 overnight, then serial dilutions of MPA (5, 10, 15, 20 and 25 µg/mL) were added; whereas for vehicle control, an equal volume of PBS was used. After 48 h, all of the cells were trypsinized and resuspended in annexin-binding buffer and stained with Alexa Fluor 488 annexin V and PI at room temperature for 15 min. Detection of apoptosis was performed by FACS.
Xenograft Assays in Nude Mice
The xenograft tumor model was performed using nude mice in accordance with current prescribed guidelines and under a protocol approved by the Institutional Animal Care and Use Committee of Hangzhou Normal University, China. Mice were bred in an SPF environment during the whole experimental period. Mice were all female and 4 to 6 wks of age at the time of inoculation, and were subcutaneously inoculated with 5 × 106 of HeLa cells. After 20 h, mice were divided into three groups and were treated with different doses of MPA or PBS (240 mg/kg body weight, n = 10; 60 mg/kg body weight, n = 11 and PBS, n = 10) (2). MPA was injected intraperitoneally (IP) for 20 consecutive days. Tumor formation was monitored through palpation. At d 30 postengraftment, mice were killed and tumors were harvested and macroscopically analyzed. Tumor tissues were fixed with 4% paraformaldehyde and embedded in paraffin for evaluation by histology or immunohistochemistry.
Paraffin-embedded tumor tissue slides were deparaffined in xylene, rehydrated in graded alcohols, and rinsed in PBS supplemented with 0.05% Tween 20. Slides were boiled in citrate acid buffer (pH 6.0) for 10 min to retrieve antigen. A 3% H2O2 treatment for 20 min at room temperature was used to block endogenous peroxidase activity. The slides were incubated in 5% milk-containing blocking solution followed by overnight incubation with either a rabbit monoclonal antibody against IMPDH2, a rabbit polyclonal antibody against p-Histone H3 or a rabbit polyclonal antibody against cleaved caspase-3, used at a final dilution of 1:500, 1:1000 or 1:3000, respectively, and then counterstained with hematoxylin according to routine procedures. As a negative control, the primary antibody was omitted; positive controls were taken from other slides that had been successfully stained before. IMPDH2, phospho-histone H3 and cleaved caspase-3 staining were scored by two independent expert observers. The numbers of mitotic cells and cleaved caspase-3-positive cells were counted in ten high-power fields. Median numbers of positive cells in each of the ten fields were calculated for each sample of the different groups using a semiquantitative assessment. Three categories were used to evaluate the percentages of apoptotic cells: <10%, mild; 10% to 50%, moderate; >50%, high. The intensity of IMPDH2 staining was presented by categories: + weak; + + moderate; + + + strong.
Statistical analysis was performed by using the nonparametric Mann-Whitney test for paired or non-paired data, or the paired t test using GraphPad InStat software as appropriate. A P value < 0.05 was considered statistically significant.
All supplementary materials are available online at https://doi.org/www.molmed.org .
MPA Is Very Effective for Inhibiting Cancer Cell Proliferation in Fast-Growing Cell Lines
MPA Inhibited Cell Proliferation, Arrested Cell Cycling and Induced Cell Apoptosis in Fast-Growing Cancer Cell Line
IMPDH2 Is a Relevant Target for MPA in Inhibiting Cancer Cell Growth
MPA Delayed Tumor Initiation, Inhibited Cancer Cell Proliferation and Induced Cell Apoptosis In Vivo
Immunohistochemical staining of tumors harvested from these mice demonstrated significant downregulation (P < 0.05) of IMPDH2 at the protein level following treatment with MPA (Figure 4C). Concomitantly, MPA inhibited tumor cell proliferation, as shown by a significant reduction (P < 0.05) of proliferating cells assessed by the percentage of p-histone H3 positive cells (Figure 4D). Furthermore, MPA treatment provokes substantial apoptosis in the tumor cell compartment, as evidenced from the significant increase in the percentage of cleaved caspase-3-positive cells (Figure 4E). These results show that MPA counteracts tumor growth elicited by a fast-growing cancer cell in vivo.
Existence of Slow-Cycling Cancer Cells Compartment in Fast-Growing Tumors
HeLa cells engineered with histone-GFPTet-on were subcutaneously engrafted in immunodeficient mice (Figure 5B). Once a small tumor was formed, histone-GFP expression was induced by doxycycline in the drinking water of the animal. Following doxycycline withdrawal, mice were killed at different time points. As shown by both FACS and immunohistochemical staining, a small population of LRCs was detected in tumors (Figures 5C,D). Thus we concluded that even fast-growing tumors harbor LRC compartment.
Slow-Cycling Cells Are Superior in Tumor Formation and Display More Resistance to MPA
Subsequently, we evaluated the relative sensitivity of LRCs and non-LRCs to MPA. Both populations were sorted and colony-forming potential was assessed in the presence or absence of MPA. Compared with the control groups, treatment with MPA (10 µg/mL) significantly inhibited the colony formation efficiency of non-LRCs but not of LRCs (where only minor effects were seen; Figure 6C). Thus, slow-cycling cancer cells when compared with fast-cycling cancer cells appear more resistant to MPA and may thus constitute the MPA-resilient reservoir in cancers.
IMPDH is a key enzyme in de novo guanine nucleotide biosynthesis and is thus a target for oncologic disease. MPA works as a potent IMPDH inhibitor that is used as an immunosuppressive drug (3). A phase I trial in patients with advanced multiple myeloma showed a positive correlation between clinical responses and depletion of the intracellular deoxyguanosine triphosphate levels by mycophenolate mofetil (MMF), the prodrug of MPA. MMF was administered up to a maximum dose of 5 g/day, which is two to three times higher than general use in organ transplantation patients, but was well tolerated in this study (19). In renal transplant patients, a tendency toward a lower risk of malignancy in MMF-treated patients versus non-MMF-treated has been reported in a large, prospectively conducted, observational cohort study (20). However, another clinical study in pancreatic cancer failed to show any beneficial effects (21). We found that only fast-cycling, but not slow-cycling, cancer cells are sensitive to the inhibitory effects of MPA. Thus, dissecting the heterogeneity of cancer may help to understand the distinct responsiveness to MPA treatment.
A unique aspect of this study is the use of a label-retaining technique that enables us to identify and isolate LRCs. Classically, nucleotide analogs, such as 5-iodo-2′-deoxyuridine and bromodeoxyuridine, are used to identify LRCs. These agents can be used to identify LRCs, but the cells are not able to be isolated for functional study (22). Thus, fluorescent-coupled nucleotide analogs were developed for identification and isolation of LRCs. However, these agents are not competent for in vivo application (23). Although nucleotide analogs were also able to label cells in vivo, subsequent isolation of LRCs from the tissue is often a challenge that hampers further functional investigation (22). Another major drawback is that most of these labeling methodologies rely on cell division to label cells. Therefore, the real quiescent cells, in fact, cannot be labeled. Consequently, the introduction of modified nucleotides into cells profoundly alters the status of the cells (24). Integrating a lentiviral-based Tet-on cell-labeling system has circumstanced these limitations. Upon induction of the GFP fluorescent protein, all the cells can be labeled without genotoxic effect. It can be used for identification and isolation of slow-cycling cells both in vitro and in vivo. In this study, we used a histone-GFP fusion protein that localized in the cell nucleus with prolonged half-life (25). Thus, this technique bears broad implications for studying cell cycling. Indeed, we found the existence of a slow-cycling cancer cell compartment within the fast growing tumors formed by HeLa cells. These cells are superior in tumor initiation and more resistant to MPA. We speculate that the existence of slow-cycling cancer cells in different patient populations may affect the ultimate responsiveness of MMF/MPA treatment (19, 20, 21).
IMPDH2 is upregulated in proliferating cells (26), including in various types of cancer cells (27,28), and exhibits a fivefold higher sensitivity to MPA compared with IMPDH1. Mechanistically, the effect of MPA appears through inhibition of its canonical target, IMPDH2. Ectopic expression of a mutated IMPDH2 (mutIMPDH2) largely nullifies the antiproliferative effects of MPA. Since IMPDH is the key enzyme in the purine nucleotide synthesis pathway, we studied the effects of a modulating purine nucleotide pool on cell growth. Increasing the nucleotide concentration by supplementation of exogenous guanosine did not have a major effect on cell growth (Supplementary Figure S4). In addition, we found, surprisingly, that supplementation of exogenous guanosine counteracts the inhibitory effect of MPA only to a minor extent (Supplementary Figure S5). Although the antiproliferative effects of MPA are mainly dependent on targeting IMPDH2, depletion of nucleotides could only explain part of its mechanism of action.
Interestingly, a recent study has demonstrated a double functionality for IMPDH in Drosophila; in stress conditions, it can act as a transcription factor that inhibits cell proliferation (29). Thus, we speculate that IMPDH2 might be a ligand-regulated transcription factor and that MPA might act as a ligand. Indeed, IMPDH2 is located predominantly in cytoplasms in normal conditions of cultured HeLa cells, but efficiently translocated into the nucleus upon MPA treatment (Supplementary Figure S6A). Consistently, a mutated IMPDH2 lacking the binding site of MPA was not able to translate into the nucleus, even with MPA treatment (Supplementary Figure S6B). These results appear to support the previous observation in Drosophila cells and our hypothesis. Furthermore, it is already known that MPA works as a ligand to activate the activity of peroxisome proliferator-activated receptors, such as PPARγ (30), a critical nuclear receptor on adipocyte differentiation. However, gene silencing of PPARα and PPARγ did not affect the sensitivity of HeLa cells to MPA treatment, excluding their potential involvement (Supplementary Figure S7). The scenario that MPA acts as a ligand for ligand-regulated transcription factors to regulate cancer cell growth is certainly interesting and therefore deserves further investigation.
In summary, this study demonstrated that through inhibiting IMPDH2, MPA was capable of constraining the growth of fast-cycling cancer cells. Using a lentiviral Tet-on cell labeling technique, we identified slow-cycling cancer cells within the fast-growing tumors that are superior in tumor initiation, but more resistant to MPA. Thus, it is very necessary to develop regimens that can effective target slow-cycling cancer cells. Combining these regimens with agents targeting fast-cycling cancer cells, such as MPA, may be a viable option in cancer therapy.
Slow-cycling cancer cells within fast-growing tumors were identified. These cells, compared with fast-cycling cells, were superior in tumor initiation and resistant to IMPDH2 inhibition by MPA. Thus, simultaneous targeting of slow- and fast-cycling cells is necessary to eradicate cancer.
The authors declare that they have no competing interests as defined by Molecular Medicine, or other interests that might be perceived to influence the results and discussion reported in this paper.
The authors thank Lifeng Ni from the Animal Care at Hangzhou Normal University, Hangzhou, China, for helping with the animal experiments. We also thank Riccardo Fodde (Department of Pathology, Erasmus Medical Center Rotterdam, The Netherland) for providing plasmids pLV.EX3D/EF1A-rtTA (M2)-dsRed-Express2 and pLV.EX2D/TRE-eGFP. Our funding included the following: support from the Daniel den Hoed Foundation for Centennial Fellowship 2014, from the Netherlands Organization for Scientific Research (NWO/ZonMw) for a VENI grant (no. 916-13-032), and from the Dutch Digestive Foundation (MLDS) for a career development grant (no. CDG 1304) to Q Pan. Funding also came from the Zhejiang Provincial Top Key Discipline of Biology (no. 2014A09-C) to K Chen and from the National Natural Science Foundation of China (no. 51272236) to X Kong.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, and provide a link to the Creative Commons license. You do not have permission under this license to share adapted material derived from this article or parts of it.
The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
To view a copy of this license, visit (https://doi.org/creativecommons.org/licenses/by-nc-nd/4.0/)